Research map / Label-efficient and robust learning / Domain adaptation and generalization
Moment Matching for Multi-Source Domain Adaptation
Xingchao Peng, Qinxun Bai, Xide Xia, Zijun Huang, Kate Saenko, Bo Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add source-source moment alignment to source-target alignment (ablation) | M3SDA-beta · Digit-Five | accuracy (%): 85.8 -> 87.7 | Improved |
| add source-source moment alignment to source-target alignment (ablation) | M3SDA-beta · Office-Caltech10 | accuracy (%): 96.2 -> 96.4 | Improved |
| add source-source moment alignment to source-target alignment (ablation) | M3SDA-beta · DomainNet | accuracy (%): 39.7 -> 42.6 | Improved |
| add source-target moment alignment to source-source alignment (ablation) | M3SDA-beta · DomainNet | accuracy (%): 34.4 -> 42.6 | Improved |
| add source-target moment alignment to source-source alignment (ablation) | M3SDA-beta · Digit-Five | accuracy (%): 81.5 -> 87.7 | Improved |
| add dynamic weight vector over source classifiers (ablation) | M3SDA · DomainNet | accuracy (%): 40.8 -> 41.5 | Improved |
| add classifier-discrepancy training (beta variant) (ablation) | M3SDA · DomainNet | accuracy (%): 41.5 -> 42.6 | Improved |
| add classifier-discrepancy training (beta variant) (ablation) | M3SDA · Digit-Five | accuracy (%): 86.13 -> 87.65 | Improved |
| add classifier-discrepancy training (beta variant) (ablation) | M3SDA · Office-Caltech10 | accuracy (%): 96.1 -> 96.4 | Improved |
| combine multiple sources into one domain for single-source DAN | DAN · DomainNet | accuracy (%): 32.9 -> 32.1 | Hurt |
| multi-source adaptation with quickdraw target | M3SDA · DomainNet quickdraw target | accuracy (%): 13.3 -> 5.2 | Hurt |
| increase number of categories from 20 to 345 | state-of-the-art DA methods · DomainNet painting->real | accuracy (%) | Hurt |
Closest papers by meaning
- Multi-Prompt Alignment for Multi-Source Unsupervised Domain Adaptation (NeurIPS 2023 · similarity 0.91)
- Co-regularized Alignment for Unsupervised Domain Adaptation (NeurIPS 2018 · similarity 0.91)
- Inferring Latent Domains for Unsupervised Deep Domain Adaptation (TPAMI 2019 · similarity 0.91)
- Incorporating Pre-Training Data Matters in Unsupervised Domain Adaptation (TPAMI 2025 · similarity 0.91)
- On Balancing Bias and Variance in Unsupervised Multi-Source-Free Domain Adaptation (ICML 2023 · similarity 0.91)
- Invertible Projection and Conditional Alignment for Multi-Source Blended-Target Domain Adaptation (AAAI 2025 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
